Publications

Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning

2026

Bin Lei, Yu Li, Prafulla Kumar Choubey, Jiaxin Zhang, Becky Xiangyu Peng, Qinyuan Ye, Kartik Narayan, Caiwen Ding, Silvio Savarese, Chien-Sheng Wu.

Download here

On the Fragility of Self-Improving Agents: Variance, Task Order, and Underspecification

2026

Qinyuan Ye, Yu Li, Yada Pruksachatkun, Jiaxin Zhang, Chien-Sheng Wu.

Download here

Best Friends, Not Forever: Evaluating Long-Horizon Persona Collapse and Behavioral Drift in AI Companions

2026

Pranav Narayanan Venkit, Akshara Prabhakar, Yu Li, Daniel Lee, Chien-Sheng Wu.

Download here

Speculate with Memory: Lossless Acceleration for LLM Agents

2026

Yu Li, Qinyuan Ye, Prafulla Kumar Choubey, Jiaxin Zhang, Chien-Sheng Wu.

Download here

Don’t Stop Early: Scalable Enterprise Deep Research with Controlled Information Flow and Evidence-Aware Termination

Published in ACL 2026 (Industry), 2026

Prafulla Kumar Choubey, Kung-Hsiang Huang, Pranav Narayanan Venkit, Jiaxin Zhang, Vaibhav Vats, Yu Li, Xiangyu Peng, Chien-Sheng Wu.

Download here

InterviewSim: A Scalable Framework for Interview-Grounded Personality Simulation

Published in COLM 2026, 2026

Yu Li, Pranav Narayanan Venkit, Yada Pruksachatkun, Chien-Sheng Wu.

Download here

Exploring the Use of VLMs for Navigation Assistance for People with Blindness and Low Vision

2026

Yu Li, Yuchen Zheng, Giles Hamilton-Fletcher, Marco Mezzavilla, Yao Wang, Sundeep Rangan, Maurizio Porfiri, Zhou Yu, John-Ross Rizzo.

Download here

The Need for a Socially-Grounded Persona Framework for User Simulation

2026

Pranav Narayanan Venkit, Yu Li, Yada Pruksachatkun, Chien-Sheng Wu.

Download here

Visual Language Model-based Food Safety Support for Persons with Blindness and Low Vision

Published in ICCV 2025 Workshops (ACVR), 2025

Ryan Hansen, Hardik Setia, Giles Hamilton-Fletcher, Aryan Jain, Zirui Liu, Mariam Zoair, Reem Aboutaleb, Qing Wen, Yu Li, John Ross Rizzo.

Download here

LIONs: An Empirically Optimized Approach to Align Language Models

Published in EMNLP 2024, 2024

Xiao Yu, Qingyang Wu, Yu Li, Zhou Yu.

Download here

From Pixels to Personas: Investigating and Modeling Self-Anthropomorphism in Human-Robot Dialogues

Published in EMNLP 2024 (Findings), 2024

Yu Li, Devamanyu Hazarika, Di Jin, Julia Hirschberg, Yang Liu.

Download here

Curriculum-Driven Edubot: A Framework for Developing Language Learning Chatbots Through Synthesizing Conversational Data

Published in SIGDIAL 2024, 2024

Yu Li, Shang Qu, Jili Shen, Shangchao Min, Zhou Yu.

Download here

User Adaptive Language Learning Chatbots with a Curriculum

Published in AIED 2023, 2023

Kun Qian, Ryan Shea, Yu Li, Luke K. Fryer, Zhou Yu.

Download here

DIONYSUS: A Pre-trained Model for Low-Resource Dialogue Summarization

Published in ACL 2023, 2023

Yu Li, Baolin Peng, Pengcheng He, Michel Galley, Zhou Yu, Jianfeng Gao.

Download here

Robots-Don’t-Cry: Understanding Falsely Anthropomorphic Utterances in Dialog Systems

Published in EMNLP 2022, 2022

David Gros, Yu Li, Zhou Yu.

Download here

Improving Conversational Recommendation Systems’ Quality with Context-Aware Item Meta Information

Published in NAACL 2022 (Findings), 2022

Bowen Yang, Cong Han, Yu Li, Lei Zuo, Zhou Yu.

Download here

Knowledge-Grounded Dialogue Generation with a Unified Knowledge Representation

Published in NAACL 2022, 2022

Yu Li, Baolin Peng, Yelong Shen, Yi Mao, Lars Liden, Zhou Yu, Jianfeng Gao.

Download here

Using Chatbots to Teach Languages

Published in L@S 2022, 2022

Yu Li, Chun-Yen Chen, Dian Yu, Sam Davidson, Ryan Hou, Xun Yuan, Yinghua Tan, Derek Pham, Zhou Yu.

Download here

Refine and Imitate: Reducing Repetition and Inconsistency in Persuasion Dialogues via Reinforcement Learning and Human Demonstration

Published in EMNLP 2021 (Findings), 2021

Weiyan Shi, Yu Li, Saurav Sahay, Zhou Yu.

Download here

LEGOEval: An Open-Source Toolkit for Dialogue System Evaluation via Crowdsourcing

Published in ACL 2021 (Demo), 2021

Yu Li, Josh Arnold, Feifan Yan, Weiyan Shi, Zhou Yu.

Download here

The R-U-A-Robot Dataset: Helping Avoid Chatbot Deception by Detecting User Questions About Human or Non-Human Identity

Published in ACL 2021, 2021

David Gros, Yu Li, Zhou Yu.

Download here

Towards Emotional Support Dialog Systems

Published in ACL 2021, 2021

Siyang Liu, Chujie Zheng, Orianna Demasi, Sahand Sabour, Yu Li, Zhou Yu, Yong Jiang, Minlie Huang.

Download here

Alternating Recurrent Dialog Model with Large-scale Pre-trained Language Models

Published in EACL 2021, 2021

Qingyang Wu, Yichi Zhang, Yu Li, Zhou Yu.

Download here

DEUX: An Attribute-Guided Framework for Sociable Recommendation Dialog Systems

2021

Yu Li, Shirley Anugrah Hayati, Weiyan Shi, Zhou Yu.

Download here

A Multi-Persona Chatbot for Hotline Counselor Training

Published in EMNLP 2020 (Findings), 2020

Orianna Demasi, Yu Li, Zhou Yu.

Download here

Gunrock 2.0: A User Adaptive Social Conversational System

Published in 3rd Proceedings of Alexa Prize, 2020

Kaihui Liang, Austin Chau, Yu Li, Xueyuan Lu, Dian Yu, Mingyang Zhou, Ishan Jain, Sam Davidson, Josh Arnold, Minh Nguyen, Zhou Yu.

Yu (Yoo) Li

Publications

Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning

On the Fragility of Self-Improving Agents: Variance, Task Order, and Underspecification

Best Friends, Not Forever: Evaluating Long-Horizon Persona Collapse and Behavioral Drift in AI Companions

Speculate with Memory: Lossless Acceleration for LLM Agents

Don’t Stop Early: Scalable Enterprise Deep Research with Controlled Information Flow and Evidence-Aware Termination

InterviewSim: A Scalable Framework for Interview-Grounded Personality Simulation

Exploring the Use of VLMs for Navigation Assistance for People with Blindness and Low Vision

The Need for a Socially-Grounded Persona Framework for User Simulation

Visual Language Model-based Food Safety Support for Persons with Blindness and Low Vision

LIONs: An Empirically Optimized Approach to Align Language Models

From Pixels to Personas: Investigating and Modeling Self-Anthropomorphism in Human-Robot Dialogues

Curriculum-Driven Edubot: A Framework for Developing Language Learning Chatbots Through Synthesizing Conversational Data

User Adaptive Language Learning Chatbots with a Curriculum

DIONYSUS: A Pre-trained Model for Low-Resource Dialogue Summarization

Robots-Don’t-Cry: Understanding Falsely Anthropomorphic Utterances in Dialog Systems

Improving Conversational Recommendation Systems’ Quality with Context-Aware Item Meta Information

Knowledge-Grounded Dialogue Generation with a Unified Knowledge Representation

Using Chatbots to Teach Languages

Refine and Imitate: Reducing Repetition and Inconsistency in Persuasion Dialogues via Reinforcement Learning and Human Demonstration

LEGOEval: An Open-Source Toolkit for Dialogue System Evaluation via Crowdsourcing

The R-U-A-Robot Dataset: Helping Avoid Chatbot Deception by Detecting User Questions About Human or Non-Human Identity

Towards Emotional Support Dialog Systems

Alternating Recurrent Dialog Model with Large-scale Pre-trained Language Models

DEUX: An Attribute-Guided Framework for Sociable Recommendation Dialog Systems

A Multi-Persona Chatbot for Hotline Counselor Training

Gunrock 2.0: A User Adaptive Social Conversational System

End-to-End Trainable Non-Collaborative Dialog System